Principal-agent software environments for multi-attribute auctions and multi-bilateral negotiations

该计划的前提是，家庭和小企业将扮演双重角色，即买家和卖家，他们将积极参与新兴市场。今天，能源的销售和购买创造了这样一个市场。可以预见，未来的智能家电将参与多个市场(例如，食品、洗衣粉、零部件和服务)。我们建议智能家电与家庭软件代理进行交互。该代理人将与代表供应商的代理人一起开展电子采购活动。这里的融合是，家庭和其他个人节点成为供应商和买家，他们以这两个角色参与市场。这种家庭和小企业代理人可能扮演的另一个角色是建立虚拟组织，以满足代理人的需求。这种组织的一个例子是由低电压和中压家庭分布式发电机组成的虚拟发电厂。代表单个发电机的代理人可能会一起工作，贡献多余的能量。他们还可能试图影响他们的委托人的行为，以便为他们提供更环保或更便宜的能源，或者增加他们的收入。这些例子显示了家庭级软件代理的潜力。 很可能，委托人会对多种属性感兴趣，而不仅仅是价格。一个原因是，这些代理人的大部分互动将是与代表邻居家庭的代理人进行的。另一个原因是人们对污染和可再生能源的兴趣增加。除了交易所的多属性性质外，最有效的机制，即拍卖，很可能需要加以修订，并辅之以谈判，以便代理人能够采用合作和个人主义的战略，而不是竞争战略。 本方案提出了逆向拍卖和多双边谈判算法的设计，并设计和实现了决策辅助工具和软件代理，为企业提供了功能丰富而又非常易用的软件环境。

This program is based on a premise that households and small businesses will play dual roles, i.e., buyers and sellers, and they will be actively participating in the emerging marketplaces. Today sales and purchases of energy create such a marketplace. One can envision that future smart appliances will participate in multiple marketplaces (e.g., food, washing powder, parts, and service). We propose that the smart appliances interact with a household software agent. This agent will engage in e-procurement activities with agents representing supply providers. The convergence here is that households and other individual nodes become suppliers and buyers and they participate in marketplaces in both roles. Another role that such household and small business agents may perform is building virtual organizations to meet the needs of the agents' principals. An example of such an organization is a virtual power plant which comprises low- and medium-voltage domestic distributed generators. Agents representing individual generators may work together and contribute excess energy. They may also attempt to influence their principals' behaviour in order to provide them with greener or cheaper energy or increase their revenues. These examples show the potential for the household-level software agents. It is likely that principals would be interested in multiple attributes rather than solely price. One reason is that most of the interactions of these agents will be with agents representing neighbourhood households. Another reason is an increase in people's interest in pollution, and renewable energy. In addition to multiattribute nature of the exchanges it is likely that the most efficient type of mechanisms, i.e., auctions would need to be revised and complemented with negotiations so that the agents can employ cooperative and individualistic rather than competitive strategies. This program proposes a design of reverse auction and multi-bilateral negotiation algorithms and design and implementation of decision aids and software agents to provide feature-rich yet very easy to use software environment.